building a frequency counter.. use 1PPS, ,or 10 MHz reference from GPS?
Hello!
I'm building a frequency counter for some amateur radio microwave
transverters to measure the local oscillators.
I am designing towards a capability of directly counting up to 100 MHz and
up to 200 MHz with a prescaler.
For a time base, I'd like to use a 1PPS pulse from a GPS, but how accurate
are they? Should I just count a 10 MHz GPS reference and use that to
generate a gate for the counter?
Yes, there will be propagation delays in dedicated logic gates, but they
don't appear to be too bad. I should be able to get down to 100 Hz or so,
which will be fine for what I need to do.
I'd rather stick with dedicated counters and gates than move to an FPGA.
Thoughts?
Thanks!
Tyler
KM3G
Shrewsbury, PA
Hi
The raw 1 pps pulse out of the typical module will have a sawtooth error on it.
How big that error is depends on the module. It could / should / might contribute
several ppb of error when used as a 1 second gate.
Bob
On Jun 23, 2023, at 2:58 PM, Tyler H via time-nuts time-nuts@lists.febo.com wrote:
Hello!
I'm building a frequency counter for some amateur radio microwave
transverters to measure the local oscillators.
I am designing towards a capability of directly counting up to 100 MHz and
up to 200 MHz with a prescaler.
For a time base, I'd like to use a 1PPS pulse from a GPS, but how accurate
are they? Should I just count a 10 MHz GPS reference and use that to
generate a gate for the counter?
Yes, there will be propagation delays in dedicated logic gates, but they
don't appear to be too bad. I should be able to get down to 100 Hz or so,
which will be fine for what I need to do.
I'd rather stick with dedicated counters and gates than move to an FPGA.
Thoughts?
Thanks!
Tyler
KM3G
Shrewsbury, PA
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe send an email to time-nuts-leave@lists.febo.com
Tyler...
Averaged over a long term, the 1 PPS from a GPS receiver is pretty accurate, but variation from one pulse to the next could be as high a 10 nanoseconds. Any GPSDO with a 10 MHz output uses the 1 PS as its reference since that is the only time signal available from a GPS receiver. The control loop for the 10 MHz ( or whatever frequency) output is designed to clean up that variation so that the variation in the 10 MHz is much less both short and long term. I have probably over-simplified this but that's the basic situation. Conclusion: The GPSDO, say a Leo Bodnar or one of the used commercial models on eBay which can usually be had for under $400, provides a very good time base for a counter.
As you mentioned microwave LO's, but only want to measure up to 200 MHz, I am guessing that you plan to measure a low frequency oscillator that is multiplied up and do some math to get the actual LO frequency. There is a very good HP app note from many years ago which talks about all of the considerations in using counters, as well as architectures. Well worth reading if you haven't already found it. Will likely save you some time and headaches, although might reduce the educational value of discovering the pitfalls of various designs yourself. There's nothing like burning your fingers to teach you which end the soldering iron is hot. Up to that point, it's only a theory 😉.
Hopefully someone on here with better search talents than mine will come up with the URL for it.
Good luck with your project!
Tom, N8ZM
-----Original Message-----
From: Tyler H via time-nuts time-nuts@lists.febo.com
Sent: Friday, June 23, 2023 2:58 PM
To: time-nuts@lists.febo.com
Cc: Tyler H tyler881@comcast.net
Subject: [time-nuts] building a frequency counter.. use 1PPS, ,or 10 MHz reference from GPS?
Hello!
I'm building a frequency counter for some amateur radio microwave
transverters to measure the local oscillators.
I am designing towards a capability of directly counting up to 100 MHz and
up to 200 MHz with a prescaler.
For a time base, I'd like to use a 1PPS pulse from a GPS, but how accurate
are they? Should I just count a 10 MHz GPS reference and use that to
generate a gate for the counter?
Yes, there will be propagation delays in dedicated logic gates, but they
don't appear to be too bad. I should be able to get down to 100 Hz or so,
which will be fine for what I need to do.
I'd rather stick with dedicated counters and gates than move to an FPGA.
Thoughts?
Thanks!
Tyler
KM3G
Shrewsbury, PA
time-nuts mailing list -- mailto:time-nuts@lists.febo.com time-nuts@lists.febo.com
To unsubscribe send an email to mailto:time-nuts-leave@lists.febo.com time-nuts-leave@lists.febo.com
On 23/06/2023 19:58, Tyler H via time-nuts wrote:
Hello!
I'm building a frequency counter for some amateur radio microwave
transverters to measure the local oscillators.
I am designing towards a capability of directly counting up to 100 MHz and
up to 200 MHz with a prescaler.
For a time base, I'd like to use a 1PPS pulse from a GPS, but how accurate
are they? Should I just count a 10 MHz GPS reference and use that to
generate a gate for the counter?
Thoughts?
Thanks!
Tyler
KM3G
Tyler,
You might like to consider this counter:
http://m0wgf.blogspot.com/2020/05/bg7tbl-fa-2-frequency-counter-review.html
This is a low-cost counter covering up to 6 GHz and could take your 10 MHz
reference as an input. I have one and it works very well, as good as the
review one.
A point about the 10 MHz, many GPS modules can produce 10 MHz but it's not as
clean a signal as they produce at 8 MHz. A 1PPS pulse from cheap GPS units is
stable to within tens of ns - so within 1 Hz at 10 MHz, perhaps?
Good luck with your project.
73,
David GM8ARV
SatSignal Software - Quality software for you
Web: https://www.satsignal.eu
Email: david-taylor@blueyonder.co.uk
Twitter: @gm8arv
Tyler,
You did not mention what accuracy is needed.
With a PPS from a modern cheap GPS receiver with a good GPS signal you can
expect a 1 s ADEV of 1e-8
Direct usage of this PPS to directly gate and count 100MHz will give you a
bit worse performance for your counter, say 3e-8, or average error of 3 Hz
per second gate time when measuring the 100MHz
With a not so good GPS signal this average error may increase to 30 Hz.
Erik.
Op za 24 jun 2023 om 03:35 schreef Tyler H via time-nuts <
time-nuts@lists.febo.com>:
Hello!
I'm building a frequency counter for some amateur radio microwave
transverters to measure the local oscillators.
I am designing towards a capability of directly counting up to 100 MHz and
up to 200 MHz with a prescaler.
For a time base, I'd like to use a 1PPS pulse from a GPS, but how accurate
are they? Should I just count a 10 MHz GPS reference and use that to
generate a gate for the counter?
Yes, there will be propagation delays in dedicated logic gates, but they
don't appear to be too bad. I should be able to get down to 100 Hz or so,
which will be fine for what I need to do.
I'd rather stick with dedicated counters and gates than move to an FPGA.
Thoughts?
Thanks!
Tyler
KM3G
Shrewsbury, PA
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe send an email to time-nuts-leave@lists.febo.com
For a time base, I'd like to use a 1PPS pulse from a GPS, but how
accurate are they?
A GPS pulse is plenty good enough for your project. You want to count
cycles of a frequency up to 100 MHz, and 100 Hz resolution "will be
fine". That's very easy because you're asking for only 6 digit, or 1 ppm
resolution and precision.
No need for a GPSDO. No need for FPGA. A microcontroller would work, but
yes, you can do this with counters and gates if you want to keep it
simple, or retro. Three 'HC390 decade counters will give you 6 digits.
You'll have to come up with something for the first digit since 100 MHz
is too fast (or just use a larger prescaler). Use the GPS/1PPS for the
start/stop gate. That's all you need for a basic 6-digit counter.
Add a 4th 'HC390 to extend to 7 or 8 digits if you want. It's only when
you get to 9 or 10 or 11 or 12 digit counters that things get
complicated. But for your stated goals, a simple GPS/1PPS 1-second gated
counter should work fine. Sounds like fun project.
/tvb
Am 2023-06-23 20:58, schrieb Tyler H via time-nuts:
I'm building a frequency counter for some amateur radio microwave
transverters to measure the local oscillators.
I am designing towards a capability of directly counting up to 100 MHz
and
up to 200 MHz with a prescaler.
For a time base, I'd like to use a 1PPS pulse from a GPS, but how
accurate
are they? Should I just count a 10 MHz GPS reference and use that to
generate a gate for the counter?
Yes, there will be propagation delays in dedicated logic gates, but
they
don't appear to be too bad. I should be able to get down to 100 Hz or
so,
which will be fine for what I need to do.
I'd rather stick with dedicated counters and gates than move to an
FPGA.
I've built a frequency counter based on a Xilinx 2C128 CPLD. It also
should
fit into a smaller 2C64, i.e. 64 FlipFlops + comb. logic. Unlike a real
FPGA,
the CPLD keeps its program data in on-chip Flash forever if wanted.
It has an SPI interface to a micro controller (BeagleBoneBlack in my
case)
The flat cable is the connection to the programmer. The Coolrunner II
family is quite old now. There are cheap Chinese programmers, I was
told.
The SMAs are RF input and 20 MHz input from a 20 MHz IQD crystal oven.
This board is made on my etching process @ home. (1.5 hours delay from
computer layout finished :-)
There is also a prettier one from commercial fabrication with the OCXO.
The 2C128 should do 100 MHz, the 2C64 should do 200 MHz just so.
I needed only 5 MHz to measure a down-converted X-band DRO.
There are on-board regulators for 3.3V (LT3080) and the 1V8 core
voltage.
74HCxxx won't work at 100-200 MHz. 74LVC might do.
regards, Gerhard DK4XP
HP app note 200. Fundamentals of Frequency Counters. Thanks!
http://leapsecond.com/hpan/an200.pdf
I miss HP.
Tyler
-----Original Message-----
From: TomH via time-nuts time-nuts@lists.febo.com
Sent: Saturday, June 24, 2023 12:01 AM
To: 'Discussion of precise time and frequency measurement' time-nuts@lists.febo.com
Cc: tholmes@woh.rr.com
Subject: [time-nuts] Re: building a frequency counter.. use 1PPS, ,or 10 MHz reference from GPS?
Tyler...
Averaged over a long term, the 1 PPS from a GPS receiver is pretty accurate, but variation from one pulse to the next could be as high a 10 nanoseconds. Any GPSDO with a 10 MHz output uses the 1 PS as its reference since that is the only time signal available from a GPS receiver. The control loop for the 10 MHz ( or whatever frequency) output is designed to clean up that variation so that the variation in the 10 MHz is much less both short and long term. I have probably over-simplified this but that's the basic situation. Conclusion: The GPSDO, say a Leo Bodnar or one of the used commercial models on eBay which can usually be had for under $400, provides a very good time base for a counter.
As you mentioned microwave LO's, but only want to measure up to 200 MHz, I am guessing that you plan to measure a low frequency oscillator that is multiplied up and do some math to get the actual LO frequency. There is a very good HP app note from many years ago which talks about all of the considerations in using counters, as well as architectures. Well worth reading if you haven't already found it. Will likely save you some time and headaches, although might reduce the educational value of discovering the pitfalls of various designs yourself. There's nothing like burning your fingers to teach you which end the soldering iron is hot. Up to that point, it's only a theory 😉.
Hopefully someone on here with better search talents than mine will come up with the URL for it.
Good luck with your project!
Tom, N8ZM
-----Original Message-----
From: Tyler H via time-nuts time-nuts@lists.febo.com
Sent: Friday, June 23, 2023 2:58 PM
To: time-nuts@lists.febo.com
Cc: Tyler H tyler881@comcast.net
Subject: [time-nuts] building a frequency counter.. use 1PPS, ,or 10 MHz reference from GPS?
Hello!
I'm building a frequency counter for some amateur radio microwave transverters to measure the local oscillators.
I am designing towards a capability of directly counting up to 100 MHz and up to 200 MHz with a prescaler.
For a time base, I'd like to use a 1PPS pulse from a GPS, but how accurate
are they? Should I just count a 10 MHz GPS reference and use that to
generate a gate for the counter?
Yes, there will be propagation delays in dedicated logic gates, but they don't appear to be too bad. I should be able to get down to 100 Hz or so, which will be fine for what I need to do.
I'd rather stick with dedicated counters and gates than move to an FPGA.
Thoughts?
Thanks!
Tyler
KM3G
Shrewsbury, PA
time-nuts mailing list -- mailto:time-nuts@lists.febo.com time-nuts@lists.febo.com To unsubscribe send an email to mailto:time-nuts-leave@lists.febo.com time-nuts-leave@lists.febo.com _______________________________________________
time-nuts mailing list -- time-nuts@lists.febo.com To unsubscribe send an email to time-nuts-leave@lists.febo.com
Hi
These days there are a lot of ways to do things. What’s right depends on a
lot of issues.
JLPCB is an outfit in China that will build a PCB and assemble it for you. They
have a selection of parts and you get to pick from their list. That makes for
some interesting choices.
A sub $2 MCU with a divider in front of it would be my choice from them.
Heading back to the GPS side of stuff, it could read out the “sawtooth correction”
message from a GPS module that sent that message. It would enable you to
do a pretty accurate counter.
Bob
On Jun 24, 2023, at 4:06 PM, Tyler H via time-nuts time-nuts@lists.febo.com wrote:
HP app note 200. Fundamentals of Frequency Counters. Thanks!
http://leapsecond.com/hpan/an200.pdf
I miss HP.
Tyler
-----Original Message-----
From: TomH via time-nuts time-nuts@lists.febo.com
Sent: Saturday, June 24, 2023 12:01 AM
To: 'Discussion of precise time and frequency measurement' time-nuts@lists.febo.com
Cc: tholmes@woh.rr.com
Subject: [time-nuts] Re: building a frequency counter.. use 1PPS, ,or 10 MHz reference from GPS?
Tyler...
Averaged over a long term, the 1 PPS from a GPS receiver is pretty accurate, but variation from one pulse to the next could be as high a 10 nanoseconds. Any GPSDO with a 10 MHz output uses the 1 PS as its reference since that is the only time signal available from a GPS receiver. The control loop for the 10 MHz ( or whatever frequency) output is designed to clean up that variation so that the variation in the 10 MHz is much less both short and long term. I have probably over-simplified this but that's the basic situation. Conclusion: The GPSDO, say a Leo Bodnar or one of the used commercial models on eBay which can usually be had for under $400, provides a very good time base for a counter.
As you mentioned microwave LO's, but only want to measure up to 200 MHz, I am guessing that you plan to measure a low frequency oscillator that is multiplied up and do some math to get the actual LO frequency. There is a very good HP app note from many years ago which talks about all of the considerations in using counters, as well as architectures. Well worth reading if you haven't already found it. Will likely save you some time and headaches, although might reduce the educational value of discovering the pitfalls of various designs yourself. There's nothing like burning your fingers to teach you which end the soldering iron is hot. Up to that point, it's only a theory 😉.
Hopefully someone on here with better search talents than mine will come up with the URL for it.
Good luck with your project!
Tom, N8ZM
-----Original Message-----
From: Tyler H via time-nuts time-nuts@lists.febo.com
Sent: Friday, June 23, 2023 2:58 PM
To: time-nuts@lists.febo.com
Cc: Tyler H tyler881@comcast.net
Subject: [time-nuts] building a frequency counter.. use 1PPS, ,or 10 MHz reference from GPS?
Hello!
I'm building a frequency counter for some amateur radio microwave transverters to measure the local oscillators.
I am designing towards a capability of directly counting up to 100 MHz and up to 200 MHz with a prescaler.
For a time base, I'd like to use a 1PPS pulse from a GPS, but how accurate
are they? Should I just count a 10 MHz GPS reference and use that to
generate a gate for the counter?
Yes, there will be propagation delays in dedicated logic gates, but they don't appear to be too bad. I should be able to get down to 100 Hz or so, which will be fine for what I need to do.
I'd rather stick with dedicated counters and gates than move to an FPGA.
Thoughts?
Thanks!
Tyler
KM3G
Shrewsbury, PA
time-nuts mailing list -- mailto:time-nuts@lists.febo.com time-nuts@lists.febo.com To unsubscribe send an email to mailto:time-nuts-leave@lists.febo.com time-nuts-leave@lists.febo.com _______________________________________________
time-nuts mailing list -- time-nuts@lists.febo.com To unsubscribe send an email to time-nuts-leave@lists.febo.com
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe send an email to time-nuts-leave@lists.febo.com